Gender Composition and Household Labour Strategies in Pre-Famine Ireland

Preprint Gender Composition and Household Labour Strategies in Pre-Famine Ireland Jane Gray Department of Sociology and National Institute for Regional and Spatial Analysis National University of Ireland, Maynooth County Kildare, Ireland Tel: +353-1-708 3596 Email: [email protected] Preprint of article to be published in The History of the Family (2006) 11, 1: 1-18. http://www.elsevier.com/wps/find/journaldescription.cws_home/620196/description 1 Abstract This paper examines the relationship between gender composition and rural household strategies in Cavan, a county in north-central Ireland, during the first half of the 19th century. I show that the ratio of adult females to males was highest in small farm households that depended for their survival on intensively deployed family labour in agriculture, flax-cultivation and spinning. By contrast, households without land or with micro-holdings relied on the income from men’s employment as agricultural labourers, supplemented by women’s work as spinners. More substantial landholders employed men as agricultural labourers. In both of the latter categories household labour strategies centred on men’s activities, with women’s work representing an important supplement, whereas in the small-farm category household labour strategies centred on a strategic balance between men’s and women’s labour input. Amongst households engaged in linen weaving the ratio of women to men was lower across all landholding categories. Differences in gender composition resulted from a complex interplay amongst household labour and inheritance strategies in a changing socio-economic environment. 2 Introduction Research on European proto-industrialization has highlighted the significance of women’s work and the gender division of labour for patterns of economic growth and development before and during the transition to factory industry. In his classic statement on the proto-industrial family economy, Medick (1976, p. 311) wrote that women’s labour often provided the “vital margin” of household subsistence. Subsequently, Gullickson (1986) argued that differences in the gender composition of the rural industrial labour force within households accounted for variations in the demo-economic consequences of proto-industrialization. According to Gray (1997, 2003, 2005), gendered differences in the interaction between household and market help to explain different and uneven patterns of development in proto-industrial regions. Goldstone (1996) suggested that differences in the extent to which women were available for work outside the household explain why, in contrast to Europe, China failed to develop a factory-based system of textile production until the 20th century. In a comparison of women’s work in three European textile manufacturing regions Maynes (2003, p. 63) concluded that, “the historically particular nature and degree of labor market participation of Europe’s young women had important implications both for the process of economic development in Europe and for the history of European age and gender relations.” The scholarship on proto-industrialization and gender has highlighted the extent to which women’s labour was simultaneously central to household prosperity and regional economic growth, and marginal insofar as women were generally confined to the most labour-intensive and poorly remunerated manufacturing tasks. Scholars have developed a number of explanations for the existence of gender divisions of labour in pre-industrial and early industrial households. These have 3 centred on the compatibility of different kinds of tasks with women’s primary responsibility for childcare (Tilly and Scott 1978; Pfister 1995), on the exclusion of women from formal social networks governed by guild regulations or hierarchies of skill (Quataert 1985; Rose 1992; Ogilvie 2004), and on women’s household ‘bargaining positions’ in changing local and global contexts (Gray 2005). But we know relatively little about how individual rural industrial households deployed women’s labour in order to meet their requirements, given the constraints and opportunities presented by gender divisions of labour. Did the significance and availability of women’s labour vary across different kinds of households, and if so, how were these variations linked to household strategies? In a classic article on the Irish linen industry, Collins (1982, p. 134) argued that, “the expanding prosperity of domestic production depended…on the paradoxical situation of a flexible household membership within the inflexibility of the parameters of production.” Those inflexible parameters included family life cycle stage and the gender division of labour between spinning (women’s work) and weaving (men’s work), “particularly as the ratio of spinning to weaving labour was not on a one to one basis” (Collins 1982, p. 133). In order to meet varying demands for men’s and women’s labour, linen manufacturing households employed itinerant spinners, relatives outside the nuclear family, journeyman weavers and servants, in addition to their adolescent sons and daughters who were able to remain at home longer than would have been possible in the absence of rural industry. This article expands on Collins’ work through a detailed examination of the surviving 1821 census records from Cavan, a county in north-central Ireland that formed part of the extended flax spinning and linen-weaving region centred on Belfast in the northeast. My analysis shows that households’ capacity to exercise flexibility 4 in their labour supply, and the proportions of male and female labour available to them, varied according to the size of their landholdings. On average, smallholders who were not engaged in weaving had higher ratios of female to male workers than either landless households or more substantial farm households. I argue that, while this was partly because smallholders did not command the resources to add or retain male labour, it also represented a distinct household economic strategy centring on labour intensive agricultural production and spinning. The analysis further shows that households engaged in weaving had lower ratios of female to male workers, irrespective of landholding size. Almost no weaving households could have produced sufficient yarn for their needs from within their own household resources. Women’s labour is more likely to have provided the vital margin of subsistence in small farm households than in weaving households. I suggest that this has important implications for how we interpret evidence on the demo-economic consequences of protoindustrialization in the Irish case. Like Collins, I assume that both gender composition and the deployment of women’s labour were subject to strategic choices and adaptations at the level of individual households. A recent special edition of this journal highlighted the fruitfulness, but also the limitations of the ‘family strategies’ concept. By focusing on strategies, “the agency of historical actors might be exaggerated, and the extent to which we can reconstruct these strategies from the resulting behavior of those actors will certainly be exaggerated” (Engelen, Kok and Paping 2004, p. 248). In particular, it is not meaningful to regard an observed behaviour as strategic if there were no alternative options. In this article I try to avoid the more serious methodological hazards associated with the household strategies approach by supplementing my data with evidence from a range of sources, and by linking the findings to trends across 5 time. However, I would also like to suggest that we can learn a good deal about household dynamics and their relationship to long-term patterns of social change by examining behaviour as though it were subject to choice, even if it did not always appear so to the actors themselves. The gender division of labour is often treated as a phenomenon governed more by normative than strategic economic considerations (see e.g. Pfister 1995, p. 139). In Ireland, so long as linen manufacturing was carried on in ‘independent’ households, men never resorted to the comparatively labour intensive and poorly remunerated task of spinning, and women rarely engaged in weaving.1 This gender division of labour created imbalances in the labour supply within households that in turn created a regional division of labour between spinning and weaving districts, and structured the overall development of the industry in Ireland (Collins 1982, Gray 1993, 1997, 2003). It broke down once the relations of production were transformed by the application of capital to the mechanization of spinning. Both women and men engaged in weaving mill-spun yarn put out by manufacturers (Collins 1997). It is important to remember that proto-industrial households could have adopted a different approach to the allocation of tasks associated with linen manufacturing. In principle, they could have allocated labour according to overall household requirements rather than by sex. Instead, they responded to the opportunities presented by the growing demand for linen goods within the parameters of the ‘traditional’ division of labour that made spinning women’s work. However, that women took up weaving after the mechanization of spinning shows that this ‘algorithm of everyday life’ (Van der Linden, quoted in Engelen, Kok and Paping 2004, p. 249) was not impervious to change when economic circumstances made it unsustainable. Thus the concept of family adaptive strategies provides a useful 6 framework for understanding variations in the gender composition of rural households. Within the parameters of a division of labour that confined women to labour-intensive, ‘drudgerous’ tasks, the significance of women’s work to the household economy varied according to landholding size. County Cavan in 1821: Data and Study Area The first comprehensive census of Irish population was carried out in 1821. Unfortunately most 19th century Irish census enumerators’ schedules were destroyed by fire in the early 20th century. However manuscript copies of the original 1821 schedules survive for fifteen parishes in County Cavan, and are held in the National Archives of Ireland (CEN 1821/1-15).2 This represents the largest cluster of surviving schedules from 1821. I collected two sets of data from these returns. The first consists of the population of households where at least one weaver was resident. The second consists of a systematic sample with a random start of every fifth household where no weaver was resident. In the analysis that follows, data from the population of weaving households are generally treated separately from the sample of non-weaving households. However, where statistical inferences are made to the entire population of the study area (weaving and non-weaving households), the weaving data are weighted by a multiplier of 0.2. <Figures 1 and 2 about here> Figures 1 and 2 show the topography of County Cavan and the location of parishes within the study area. The county was incorporated to Ireland’s growing domestic linen industry in three waves. During the first half of the 18th century the growth of population in the linen districts around Belfast led to increased demand for food, milch cattle and yarn, which in turn stimulated improved agricultural production 7 and an increase in spinning for the market throughout south Ulster (Crawford 1975, p.245). Weaving was at first confined to Protestant settlers along the northern border of County Cavan, but after 1745 it spread throughout the religious and social hierarchy in response both to widespread crop failure and cattle losses, and to increased demand for Irish linens in Britain and the American colonies. Irish linens were exposed for sale by weavers and small manufacturers through a network of ‘brown’ linen markets, so-called because the linens were unbleached. There they were purchased by drapers who sold them on to the bleacher-merchants that dominated the Irish linen trade by the second half of the 18th century. In Cavan, the most important market for linens was at Cootehill, in the east of the county. There was a secondary market at Killashandra, in the west, with two smaller markets (Arvagh and Ballynagh) to its south. The 1780s and 90s saw unprecedented increases in Irish linen exports. Sales at Cavan’s linen markets more than doubled in this period (Crawford 1994, p. 78), which also saw the beginning of a new wave of agricultural commercialisation in response to high wartime prices for corn (O’Neill 1984, pp. 7278). Linen sales in Cavan began to decline in the early decades of the 19th century, although the number of weavers attending the markets does not seem to have changed (see Crawford 1988, p. 47). There is some evidence that the volume of output increased again in the 1830s and 1840s (Collins 1982, p. 142). Cavan’s markets accounted for about 5% of the total value of linens sold at brown linen markets in Ireland in 1821. Unfortunately, the area covered by the surviving census schedules does not include either of the main linen market towns, but it does include the minor market of Ballynagh, and eight of the fifteen parishes lie within twenty kilometres of either Cootehill or Killashandra. My sample implies a proportion of 8.6 percent of households engaged in weaving within the study area. 8 The proportion is 8.4 percent if the population of weaving households in the census manuscripts is calculated against the published totals from the early 19th century. Just over 50 percent of all households had at least one spinner resident, but this estimate is unreliable as the proportion of spinners in the population is dependent on the proportion of women for whom an occupation was recorded and this, in turn, varied considerably across parishes within the study area (see Gray 1999, 2005). Spinning was an everyday part of normal household duties for all women but I will argue, the importance of spinning to the family economy varied across different kinds of households. Unfortunately we do not have data on yarn sales in Cavan comparable to the data on linen sales, but observations by contemporaries indicate that the county was a major supplier of yarn from the beginning of the 18th century through the early decades of the 19th century. Using econometric models, McGregor (1992) showed that domestic industry was associated with a reduction of inequality in the distribution of landholdings in Ireland in the 1840s, without having any discernible effect on median landholding size. This is partly consistent with the model of the proto-industrial family economy developed by Medick (1981a, pp. 48-49). The availability of income from rural industry lifted the restrictions on household formation inherent in the peasant economy, permitting families to survive on smaller landholdings. At the same time the desire to ensure an adequate basis for subsistence meant that rural industrial producers were willing to pay exorbitant rents just to maintain a grip on the land. As a result, proto-industrial regions were often characterized by the rapid expansion of smallholders (Medick 1981b, p. 83). In Ireland cottiers in the vicinity of major linen markets were able to outbid tenant farmers for leases, leading to an overall reduction in landholding size (Crawford 1975, p. 255).3 Elsewhere, however, farmers retained 9 relations – increased productivity simply led to increased rents – the Irish had no incentive to postpone marriage.8 Thus Connell represented early marriage as an ‘inferior good’ (O’Grada 1994, p. 7) leading to the extraordinary growth in Irish population that occurred from the middle of the18th century. Research carried out in the 1980s and 1990s showed that at least some aspects of Connell’s argument are clearly wrong. Most importantly, there is little evidence to support the thesis that the Irish married at an unusually young age (Guinnane 1997, p. 82). There is also some evidence of an increase in average age at marriage in the 1820s and 1830s (O’Grada 1994, pp. 69-74; O’Neill 1984, p. 181), suggesting that Irish nuptiality did in fact respond to changing economic circumstances. Some research supported the hypothesis of a positive relationship between participation in rural industry and propensity to marry but the findings are ambivalent from the perspective of proto-industrialization theory (see the discussions in Kriedte, Medick and Schlumbohm 1997 and Gray 2005). In an analysis of aggregate, county-level data from the 1841 census, Almquist (1979, p. 711) found that spinning was significantly positively associated with high young female nuptiality. However, he found no association between weaving and young female nuptiality, although weaving was associated with young male nuptiality. Moreover, spinning and weaving were not significantly positively correlated, reflecting the extent to which large parts of rural Ireland were incorporated to the linen industry only through spinning. In a more sophisticated analysis of 1841 census data at the smaller territorial level of barony, Mokyr (1985, p. 55) found that cottage industry affected male propensity to marry, but not that of females. However, Mokyr measured ‘cottage industry’ as the proportion or rural men and women employed in ‘occupations ministering to 16 clothing,’ thus obscuring the different effects of spinning and weaving suggested by Almquist’s analysis. Unfortunately, the 1821 census did not collect direct information about marriage or marital status (other than relationship to the head of household). However, O’Neill’s (1984, pp. 177-186) analysis of data from the surviving 1841 manuscripts for the parish of Killashandra (see Map 2) revealed an average age at marriage of between 25 and 26 years for men, and about 22 years for women. These figures are low compared to estimates for Ireland as whole, and Guinnane (1997, p. 82) suggested this might be explained by the availability of earnings from textile production in Killashandra. Unfortunately, O’Neill did not make separate calculations of age at marriage for weavers or spinners. He did, however, calculate that on average labourers’ wives married somewhat later - at 22.3 years - than farmers’ wives - 21.7 years. The opposite was true for their husbands: on average labourers were 24.5 years old at first marriage, while farmers were 26 years old. In O’Neill’s data the youngest mean ages at first marriage were found amongst those farmers and their wives that he defined as ‘middling’ – holding from 13-25 Irish acres.9 Husbands in this category married at about 25.6 years compared to 27.1 years amongst smaller farmers. Their wives’ ages averaged 21.6 years compared to 23.8 years amongst smaller landholders. While direct comparison cannot be made with the 1821 data, O’Neill’s findings do suggest that the opportunity for marriage declined as landholding size decreased. They are consistent with the image of a landholding system characterized by widespread subdivision that began to be resisted as the minimum viable holding size approached. According to the mid-century parliamentary enquiry known as the Devon Commission, eight acres was the minimum holding size required to “support a family in comfort” (quoted in McGregor 17 1992, p. 480), and this corresponds fairly well to the modal distribution of farm size in County Cavan.10 For farmers who wished to avoid subdividing their land, training their sons to be weavers offered an alternative way of providing them with an inheritance. In 1821 it was observed of Raphoe, in County Donegal, that twenty years previously “It was usual…for the farming heads of families to put their younger sons into this employment [weaving], many of whom were enabled by proper economy and application in a few years to purchase farms, but this practice has been long discontinued” (quoted in Crawford 1994, p. 51). It is plausible that a similar strategy persisted in County Cavan, given its closer proximity to Ulster’s core weaving zone. Amongst households in the study area with resident adult sons, 12 percent had at least one son engaged in weaving on medium sized holdings, compared to 6 percent of landless households and 8 percent of households on smallholdings. Heads of household who were weavers were a little younger on average than either farmers or labourers.11 Thus the medium-holding households in the Cavan study area may have had more adult sons resident due to a strategy of delaying or avoiding inheritance. However, they also had a higher demand for male labour, as indicated by the greater proportion of such households with male servants present. This can be attributed to the system of spade husbandry that characterized agriculture throughout most of Ireland before the Famine. The most commonly cultivated crops in Cavan were oats, followed by potatoes and flax. According to O’Neill (1984, p. 86), the prevailing system of rotation in the early 19th century was two crops of potatoes, followed by flax or oats, then three crops of oats followed by grass or potatoes. The continuous cropping regime, where the land was never put to grass, was characteristic of small 18 farms without livestock. Because agriculture was carried out by spade husbandry, tillage in pre-Famine Ireland created diseconomies of scale. According to McGregor (1992, p. 480), “The optimum size of holding fell with increases in tillage intensity…As farm size fell, production became more labor-intensive. Thus small farms tended to substitute labor for capital in production and tillage for livestock products in outputs.” On farms without livestock, considerable labour was expended in the collection of manure – from the roads, wasteland and seashore. Data from the 1841 census show that the average number of cattle and pigs per farm increased with farm size. However, according to calculations by O’Neill (1984, Table 2.8, p. 91), the proportion of land given over to livestock followed a curvilinear pattern, with ‘middling’ farms having lower mean numbers of animals per acre than either the smallest or largest farms. This implies both that the amount of land devoted to cereal tillage was greater in the middle ranges, and that the demand for male labour was highest on these farms. Even allowing for the fact that medium-sized farms tended to have larger numbers of adult sons and male servants resident, they could not have met their male labour requirements from within their own households. According to Sir Charles Coote, it took “but twelve men to dig an acre in a day’s work” (quoted in O’Neill 1984, p. 87). Farmers in County Cavan had two principal options for adding to their male labour supply at times of heightened demand. Under the cottier system they could provide landless families with a cottage and small plot for potato cultivation, in exchange for labour on the land. Cot-takes sometimes also included access to grazing land for a cow, or land for the cultivation of flax. Manufacturing farmers sometimes also employed weavers under the cottier system, a practice famously described, vilified and almost certainly exaggerated in its significance by Coote (1802, p. 41). 12 19 An alternative means of adding male labour was provided by the kinship or community based cooperative work-team known as a “meitheal.” It seems likely that smallholders were more likely to rely on the meitheal than more substantial farmers (Slater and McDonough 1994, p. 88). In the Cavan study area, we have already seen that those parishes nearest to the major linen markets were characterized by relatively evenly distributed smallholdings, and fewer landless households. The proportions of households headed by labourers were also smaller in these parishes. If middling farm households were more likely to retain adult sons at home and to hire male servants in order to contribute to the high labour demands associated with tillage agriculture under spade husbandry, why did smallholders strive to hire female servants and to retain adult daughters at home? The most plausible explanation is that the relative importance of female labour to the household economy increased as landholding size approached the “comfortable minimum” described by the Devon Commission. Women and children were more likely to have been responsible for drudgerous tasks such as the collection of manure (McGregor 1992). In addition, the income from flax cultivation and spinning is likely to have increased in significance. Flax cultivation drew on large quantities of female labour (Crawford 1991, Gray 2003), even though flax tended to be grown in garden plots of about an eighth of an acre – that is, rarely in quantities exceeding the spinning capacity of individual households (Smyth 1988). Arthur Young’s estimates of the expense of an acre of land under flax, compiled at ten locations across the northern linen counties (but not, unfortunately, in County Cavan), give an indication of the allocation of men’s and women’s labour in flax cultivation. Women and children were responsible for removing stones and weeds from the fields, and for pulling the flax when it was ready for harvest. If the seed was to be saved, men did so in a process called ‘rippling.’ 20 Men were also responsible for laying the flax in water to be ‘retted’ – literally rotted in order to break up the fibres – and for taking it out once that process was complete. Women spread the flax onto grass to dry out. Men and women then lifted the flax and carried it home (with the assistance of a horse and cart) where it was often subjected to further drying by a turf fire. The flax fibres were then broken up by ‘beetling,’ and ‘scutching.’ These processes essentially involved beating the flax stalks to remove the outer layers and to break the remaining fibres into strands. Flax could be brought to a scutch mill, but in many places working groups of women did the scutching, in a practice similar to the ‘meitheal’ described earlier. Young’s estimates imply a ratio of between three and four woman days to each man day required in cultivating flax, from weeding and stoning through to beetling and scutching. After scutching the flax fibres were straightened by ‘hackling’ (a sort of combing) in preparation for spinning. Specialist male ‘flax dressers’ could be employed to do this, but spinners frequently did the job themselves (see the discussion in Gray 2003).13 <Table 4 about here> Unfortunately the 1821 census does not provide very reliable evidence on the prevalence of spinning in different households or parishes. Table 4 shows the extent to which adult women were identified as spinners by landholding category in the Cavan study area. In both non-artisan and weaving households, the proportion of adult females identified as spinners was greatest amongst those who were landless, even though the mean number of adult females was lowest in this group. In weaving households, the highest mean number of spinners was observed amongst those with mid-sized holdings, in contrast to non-artisan households where the number of adult females identified as spinners did not increase greatly with landholding size. These findings have to be treated with great caution because the proportion of women for 21 whom an occupation was recorded in the 1821 census manuscripts, and therefore the proportion identified as spinners, is highly skewed towards a cluster of parishes in the southeast of the study area that were characterized by relatively high proportions of landless labourers on poor quality land (see the discussion in Gray 1999, 2005). A somewhat different picture is provided by published data from the 1841 census. The percentage of women ‘ministering to clothing’ by parish in 1841 – a reasonable proxy for spinning - is negatively correlated with the proportion of landless households in 1821 (Spearman’s rho = -.768, p. < .000). It is positively associated with the proportion of households with mid-sized landholdings (Spearman’s rho = .615, p. < .007) and also – albeit more weakly - with the proportion of smallholding households (Spearman’s rho = .442, p. < .05). There is a negative association between spinning in 1841 and distance from the nearest linen market amongst parishes within the study area (Spearman’s rho = -.559, p. < .015). Even assuming that the 1841 census recorded women’s occupations more reliably than that of 1821, aggregate data at the level of parishes can tell us little about the relative importance of earnings from spinning in different kinds of households. Moreover, by 1841, daily earnings from hand spinning had halved due to competition from the spinning mills. For landless households this loss of income compounded the distress caused by falling wages in the context of increased population and declining agricultural prices (see O’Neill 1984, especially 108-115). In this context it is interesting to note that labourers’ daughters were much less likely than farmers’ daughters to be recorded as spinners in the 1841 census manuscripts for Killashandra (O’Neill 1984, pp. 148-149). In sum, census data at both household and individual level suggest that, while spinning may have been an important source of income for landless and labouring households in 1821, by 1841 women in such households were 22 much less likely to record spinning as an occupation in the context of rapidly falling prices. There is little evidence of variation in the proportion of female labour devoted to spinning between smallholding and medium-holding households – if anything women in the latter category seem to have been more likely to record spinning as an occupation. However, spinning was more prevalent in parishes adjacent to the major linen markets characterized by a smallholding land pattern and few landless labourers. The evidence presented above is therefore not inconsistent with the argument that, in the context of labour-intensive agricultural production and the opportunity to earn additional income from spinning, the relative importance of women’s labour to the small-farm economy increased as landholding size decreased within the Cavan study area. More substantial farmers sought to maximize male labour by retaining adult sons at home and by employing day labourers and cottiers. However, by themselves, the census records provide only a partial understanding of the socioeconomic environment within which the members of rural industrial households strove to make a living. In the following section, in order to obtain a richer picture of the political economy of County Cavan, data from the 1821 census are compared with the “Statistical Memoirs” compiled by army officers employed by the Ordnance Survey in 1835. Local case studies: Drumlumman, and Drung and Larah There are three parishes for which both census manuscripts and statistical memoirs are available, namely Drumlumman in the southwestern part of the county, and Drung and Larah in the east (see Map 2). Because the latter two parishes were enumerated together in the 1821 census (see note 2) they will be treated as one parish in this discussion. Lieutenant Andrew Beatty compiled the memoir for Drumlumman 23 (reproduced in Day and McWilliams 1998, pp. 12-18), while Lieutenant P. Taylor compiled the memoirs for both Drung (Day and McWilliams 1998, pp. 19-24) and Larah (Day and McWilliams 1998, pp. 39-46). At first glance, the census data suggest that the parishes were quite similar in social structure. The mean landholding size in 1821 was between 9 and 10 acres, and just fewer than 10 percent of households had resident weavers in each case. However, in Drumlumman, 47 percent of households held either no land or micro holdings of less than 3 acres, compared to 28 percent in Drung and Larah. Moreover, the Statistical Memoirs make it clear that small landholdings in Drumlumman resulted from the practise of sub-letting, whereas in Drung and Larah they resulted from subdivision. This difference was at least partly due to Drung and Larah’s relative proximity to the long established linen market at Cootehill, and yarn market at Cavan town.14 It may also have been due to inattentive estate management in Drumlumman. According to Beatty, “No gentleman possessing any property in the parish resides in it, nor do any of their agents. The agents are generally paid by a percentage on the rent collected “ (Day and McWilliams 1998, p. 17). Under these circumstances landlords may have been less zealous about removing “middlemen” at the expiration of leases (see note 3).15 In Drumlumman, farmers who paid their rents wholly in money rented holdings of between 10 and 20 acres under leases of lives. They sublet portions of land, from 3 to 5 acres, to under-tenants who paid their rents partly in labour. Farmers relied principally on spade husbandry to cultivate potatoes, oats and sometimes wheat, but horse-drawn ploughs were used by “the better description of farmers.” The farmers burned limestone on their own land for use as manure. Oats, cattle and pigs were bought by jobbers at local fairs and markets who transported 24 them to the port of Drogheda where they were sold for export to England. Beatty’s account suggests that little milling was carried on within the parish. Forty-three percent of weavers in Drumlumman were heads of households in 1821, most of which had no land attached. Most of the remainder (nearly 50%) were employees or lodgers in farm households; just 7 percent were sons. This suggests that weaving was mainly carried on under the journeyman system whereby farmers put out yarn either to cottiers on their land, or to boarders in their own houses. Such weavers were probably also employed seasonally as agricultural labourers. According to the Statistical Memoir, “Weaving coarse linens is practised by the men in the summertime, when their farming does not occupy all their time. The women are employed in hand-spinning but of later years this [has] not given any fair remuneration for labour” (Day and McWilliams 1998, p. 16). Forty-four percent of women in Drumlumman reported an occupation ‘ministering to clothing’ in 1841. Curiously, Beatty reported that “The produce of the loom is consumed on the spot.” This may mean that linens were manufactured for local consumption only, but it seems more likely that they were purchased by local jobbers (see the discussion in Crawford 1988, p. 50). The evidence suggests that there must have been a considerable over-supply of agricultural labourers in this parish in the 1830s, and indeed Beatty provides a vivid description of their poverty: The food of the poorer class consists chiefly of potatoes and buttermilk. In winter they are seldom able to get milk and they substitute a drink made of onions boiled in water. Some of them assert that they are in such poverty at times that they are unable to purchase salt. This seems almost incredible. 25 Acknowledgements The data on which the analysis in this article is based were originally collected with the assistance of the Woodrow Wilson Foundation Women’s Studies Programme, the Economic and Social Research Institute of Ireland, and an Indiana University South Bend Summer Faculty Fellowship. I would like to thank James Keenan for creating the maps, which were funded by an NUI Maynooth Library Publication Grant. Lexington Books generously gave permission to reproduce them. 32 Endnotes 1 Women are more likely to have engaged occasionally in weaving than men are to have engaged in spinning. According to Crawford (1991, p. 260) it is “probable” that women wove during busy times. McKernan (1995) has shown that efforts to encourage women to weave using the fly shuttle had some success in County Armagh in the context of male labour shortages during the Napoleonic wars. Nevertheless, the preponderance of evidence shows that Irish women did not engage in linen weaving to any significant extent until after the mechanization of spinning. 2 Strictly speaking, these records cover sixteen civil parishes. However, in the 1821 census the parishes of Drung and Larah were enumerated as though they were one, probably because a small part of Larah is surrounded by Drung (see Map 1). 3 Some landlords were happy to facilitate subdivision in this way, because they believed that otherwise the rental value of their estates was accruing to farmermiddlemen, rather than themselves (see Coote 1802). 4 Cunningham (1960) concluded that the landholdings in the 1821 census manuscripts for the parish of Lavey were recorded in Cunningham acres. However, according to O’Neill (1984, p. 91, Table 2.7), Irish acres were the standard measure in County Cavan. 5 This pattern is consistent with O’Neill’s (1984, pp. 171-177) analysis of child/woman ratios in the surviving 1841 census manuscripts for the parish of Killashandra, in County Cavan. He found that, after the first three years of marriage, farming women consistently had more children under age five living in their households than labouring women. 33 6 Here, “inmate” refers to a resident not identified either as kin of the head of household, or as a servant. It includes residents without any identifiable relationship to the household head. 7 In the parish of Kildrumsherdan in 1835, male servants earned between 4 and 6 pounds per year, and female servants between 2 and 4 pounds per year in addition to their board and lodging (Day and McWilliams 1938, p. 37). 8 However, in parts of the province of Ulster, including County Cavan, tenants benefited from the customary observation of ‘tenant right,’ which permitted them to sell their interest in their holdings to another tenant, and thereby to obtain a return from capital investment. 9 There were thus somewhat larger, on average, than the mid-sized landholdings in this study. 10 That is, about 6 Cunningham acres, or 5 Irish acres, assuming that the observation referred to statute acres. 11 The average age of heads of households who were weavers was 40.53, compared to 45.39 amongst farmers (95% C.I. 44.70-46.08) and 43.16 amongst labourers (95% C.I. 42.11-44.21). 12 The term ‘cottier’ has been used generally by Irish historians to refer to those cottier-labourers who obtained access to land in full or part exchange for labour, and to distinguish them from small tenant farmers and day labourers. Beames (1975) noted that the term was used rather loosely amongst contemporaries. In some parts of the country ‘cottier’ referred to any smallholder, and in others simply to somebody who lived in a cabin, irrespective of the size of their holding or their occupation. Confusingly, Beames identifies Cavan as one of the places where the term ‘cottier’ 34 was used in the latter sense, but it is clear from Coote’s account that he was referring to the cottier-labourer system. 13 Young’s estimates must be treated with some caution, since they did not include the labour costs of ploughing and sowing, and because few growers cultivated so much as an acre. Nonetheless, his estimates are of the same order as those Mendels (1981, pp. 134-135) provided for the cultivation of flax in Flanders. Young (1892, pp. 138-139) did speak to a weaver on the Ards peninsula in East Ulster who grew the more usual ‘peck’s sowing.’ This man did not give a complete breakdown of labour costs by gender, but his account does imply a more even distribution of male and female labour, given that he had his flax rippled (which increased the male labour input), and brought it to a mill to be scutched (which decreased the female labour input). It is not entirely clear why Irish flax sowers rarely saved their seed, relying instead on imported flaxseed from North America. In the fine weaving districts around Belfast the flax stalks were pulled before the seed had ripened in order to produce fine fibres. Elsewhere, according to Gill (1925, p. 34), given the small scale of production in Ireland, and the ready availability of American flaxseed, it may not have been worth the growers’ while to set some of the stalks aside to dry before processing them for spinning coarse yarn. At Waringstown, County Down, Young (1892, p. 132) was informed that “Very few save their seed; but this more than usual, owing to the import from America falling off.” See also the discussion in Hood (2003). 14 At their mid-point, Drung and Larah are 11 kilometres from Cootehill. Drumlumman is 19 kilometres from Killashandra at its midpoint, although the satellite market towns of Arva and Ballynagh are nearer. However, these linen markets in the west of the county were less well established than those in the north35 east, and so the processes associated with land subdivision had had less time to develop. 15 It should be noted, however, that Beatty also referred to the activities of Lord Farnham, a famously improving landlord, in establishing a school and in contributing to the cost of constructing farm buildings, in parts of the parish. 16 The regional concentration of the linen industry around Belfast associated with the mechanization of spinning led to an increase in emigration from the western and southern counties of Ulster in the 1830s. See Collins 1982, esp. p. 140 and O’Grada 1994, p. 76. 17 The latter could be determined by a whole range of factors, including the availability of waste land, inattentive estate management, or the practise of holding land in common. 36 References Almquist, Eric L. 1979. “Pre-Famine Ireland and the theory of European protoindustrialisation: evidence from the 1841 census.” Journal of Economic History 39: 699-718. Beames, Michael. 1975. “Cottiers and Conacre in Pre-Famine Ireland.” Journal of Peasant Studies 2, 3: 352-354. Cohen, Marilyn. 1990. “Peasant Differentiation and Proto-Industrialisation in the Ulster Countryside: Tullylish 1690-1825.” Journal of Peasant Studies 17: 413-432. Collins, Brenda. 1982. “Proto-Industrialization and Pre-Famine Emigration.” Social History 7: 127-146. Connell, Kenneth H. 1950. The Population of Ireland, 1750-1845. Clarendon: Oxford University Press. Coote, Sir Charles. 1802. A Statistical Survey of the County of Cavan. Dublin: The Dublin Society. Crawford, Margaret. 1995. “Poverty and the Famine in County Cavan.” Pp. 139-158 in Cavan: Essays on the History of an Irish County, edited by Raymond Gillespie. Dublin: Irish Academic Press. Crawford, William H. 1994. The Handloom Weavers and the Ulster Linen Industry. Belfast: Ulster Historical Foundation. (First published 1972 as Domestic Industry in Ireland: The Experience of the Linen Industry. Dublin: Gill and Macmillan). ______. 1991. “Women in the Domestic Linen Industry.” Pp. 256-264 in Women in Early Modern Ireland, edited by Margaret McCurtin and Mary O’Dowd. Dublin: Wolfhound Press. 37 ______. 1988. “The Evolution of the Linen Trade in Ulster Before Industrialization.” Irish Economic and Social History 15: 32-53. ______. 1975. “Economy and Society in South Ulster in the Eighteenth Century.” Clogher Record 8: 241-258. Day, Angelique and Patrick McWilliams, eds. 1998. Ordnance Survey Memoirs of Ireland, Volume 40. Counties of South Ulster 1834-8. Cavan, Leitrim, Louth, Monaghan and Sligo. Belfast: Institute of Irish Studies. De Vries, Jan. 1993. “Between purchasing power and the world of goods: understanding the household economy in early modern Europe.” Pp. 85-132 in Consumption and the World of Goods, edited by John Brewer and Roy Porter. London: Routledge. Engelen, Theo, Jan Kok and Richard Paping. 2004. “The family strategies concept: An evaluation of four empirical case studies.” History of the Family 9: 239251. Gill, Conrad. 1925. The Rise of the Irish Linen Industry. Oxford: Oxford University Press. Goldstone, Jack. 1996. “Gender, Work, and Culture: Why the Industrial Revolution Came Early to England But Late to China.” Sociological Perspectives 39: 121. Gray, Jane. 2005. Spinning the Threads of Uneven Development: Gender and Industrialization in Ireland during the Long Eighteenth Century. Lanham, Maryland: Lexington Books. ______. 2003. “The Irish, Scottish and Flemish Linen Industries During the Long Eighteenth Century.” Pp. 159-186 in The European Linen Linen Industry in 38 Historical Perspective. Pasold Studies in Textile History 13, edited by Brenda Collins and Philip Ollerenshaw. Oxford: Oxford University Press. ______. 1999. “Spinners and Spinning in the Political Economy of Pre-Famine Ireland: Evidence from County Cavan.” Pp. 161-184 in Reclaiming Gender: Transgressive Identities in Modern Ireland, edited by Marilyn Cohen and Nancy J. Curtin. New York: St. Martin’s Press. ______. 1997. “The Irish and Scottish Linen Industries in the Eighteenth Century: An Incorporated Comparison.” Pp. 37-70 in The Warp of Ulster’s Past: Interdisciplinary Perspectives on the Irish Linen Industry, 1700-1920, edited by Marilyn Cohen. New York: St. Martin’s Press. ______. 1993. “Rural Industry and Uneven Development: The Significance of Gender in the Irish Linen Industry.” Journal of Peasant Studies 20: 590-611. Guinnane, Timothy. 1997. The Vanishing Irish: Households, Migration and the Rural Economy in Ireland, 1850-1914. Princeton: Princeton University Press. Gullickson, Gay. 1986. Spinners and Weavers of Auffay. Cambridge: Cambridge University Press. Harris, Ruth-Ann. 1994. The nearest place that wasn’t Ireland: early nineteenthcentury Irish labor migration. (Ames, Iowa: University of Iowa Press.) House of Commons. 1843. Vol. 24. Report of the Commissioners Appointed to Take the Census of Ireland, for the Year 1841. (Reprinted in 1968, Irish University Press Series of British Parliamentary Papers. Population 2. Shannon: Ireland). McGregor, Pat. 1992. “The Labor Market and the Distribution of Landholdings in Pre-Famine Ireland.” Explorations in Economic History 29: 477-493. 39 McKernan, Anne. 1995. “War, Gender and Industrial Innovation: Recruiting Women Weavers in Early Nineteenth Century Ireland.” Journal of Social History 28: 109-124. Maynes, Mary Jo. 2004. “Gender, Labor and Globalization in Historical Perspective. European Spinsters in the International Textile Industry, 1750-1900.” Journal of Women’s History 15:47-66. Medick, Hans. 1981a. “The proto-industrial family economy.” Pp. 38-73 in Industrialization before Industrialization. Rural Industry in the Genesis of Capitalism by Peter Kriedte, Hans Medick and Jurgen Schlumbohm. Translated by Beate Schempp. Cambridge: Cambridge University Press. ______. 1981b. “The structures and function of population-development under the proto-industrial system.” Pp. 74-93 in Industrialization before Industrialization. Rural Industry in the Genesis of Capitalism by Peter Kriedte, Hans Medick and Jurgen Schlumbohm. Translated by Beate Schempp. Cambridge: Cambridge University Press. ______. 1976. “The proto-industrial family economy: the structural functions of household and family during the transition from peasant society to industrial capitalism.” Social History 3: 291-315. Mendels, Franklin. 1981. Industrialization and Population Pressure in EighteenthCentury Flanders. Dissertations in European Economic History. Originally presented as the author’s thesis, University of Wisconsin, Madison, Wisconsin, 1969. New York: Arno Press. Mokyr, Joel. 1985. Why Ireland Starved: A Quantitative and Analytical History of the Irish Economy, 1800-1850. London: Allen and Unwin. 40 National Archives of Ireland. CEN 1821/1-15. Shelf Numbers 2/436/4-7. Copies of 1821 census enumerators’ schedules, County Cavan, various parishes. Ó’Gráda, Cormac. 1994. Ireland: A New Economic History, 1780-1939. Oxford: Clarendon Press. O’Neill, Kevin. 1984. Family and Farm in Pre-Famine Ireland: The Parish of Killashandra. Madison: University of Wisconsin Press. Ogilvie, Sheilagh. 2004. “How Does Social Capital Affect Women: Guilds and Communities in Early Modern Germany.” American Historical Review 109: 325-359. Pfister, Ulrich. 1996. “A General Model of Proto-Industrial Growth.” Pp. 73-92 in Proto-Industrialization: Recent Research and New Perspectives in Memory of Franklin Mendels, edited by Rene Leboutte. Geneve: Librairie Droz. ______. 1995. “The Protoindustrial Household Economy: Toward a Formal Analysis.” Pp. 114-145 in The European Peasant Family and Society: Historical Studies, edited by Richard L. Rudolph. Liverpool: Liverpool University Press. Quaetaert, Jean H. 1985. “The Shaping of Women’s Work in Manufacturing: Guilds, Households and the State in Central Europe, 1648-1870.” American Historical Review 90: 1122-1149. Rose, Sonya. 1992. Limited Livelihoods: Gender and Class in Nineteenth-Century England. Berkeley: University of California Press. Schlumbohm, Jurgen. 1996. “Proto-industrialization as a research strategy and a historical period – a balance sheet.” Pp. 12-22 in European protoindustrialization, edited by Sheilagh C. Ogilvie and Markus Cerman. Cambridge: Cambridge University Press. 41 Table 1. Ratio of Adult Females to Males in Non-Artisan Households by Landholding Category and Family Life Cycle Stage Landholding Category Landless/Micro Small Holding Medium Holding Family Lifecycle Stage1 N % Sex Ratio N1 % Sex Ratio N % Sex Ratio 1.Beginning229 5.6 1.24 43 7.7 1.12 16 3.9 1.18 2.Growing 317 61.7 1.15*** 310 55.2 1.26 187 45.3 1.26 3.Mature 116 22.6 1.18 150 26.7 1.33 168 40.7 1.17* 4.Advanced 34 6.6 1.21 52 9.3 1.32 37 9.0 0.89*** 5.Empty Nest 18 3.5 N/A 6 1.1 N/A 5 1.2 N/A All Stages 514 100 1.16*** 562 100 1.27 413 100 1.19*** Difference from smallholders * p. <.10; *** p. <.01. Mann-Whitney Test. 1. Family lifecycle stage could not be determined for one household 2. Beginning – Wife less than 50 years, no children; Growing – Children less than 15 years only; Mature – Children less than 15 years and 15 years or more; Advanced – Children 15 years or more only; Empty nest – Wife 50 years or more, no children. Table 2. Non-artisan households in each landholding category, by presence of different kinds of adult residents % Households with one or more adult (15+ years) Females Landholding Size N Daughters Servants Relatives Inmates Landless 514 18.9 2.7 5.9 7.8 Smallholding 562 25.3 12.5 6.2 6.0 Mid-holding 413 33.4 27.8 6.8 10.7 Males Landholding Size N Sons Servants Relatives Inmates Landless 514 20.6 1.8 2.1 3.5 Smallholding 562 25.5 4.8 2.0 2.5 Mid-Holding 413 40.4 23.2 4.6 6.8 Source: Cavan Sample. Table 3. Weaving households in each landholding category, by presence of different kinds of adult residents % Households with one or more adult (15+ years) Females Landholding Category n Daughters Servants Relatives Inmates Landless/Micro 366 15.8 4.9 4.9 9.3 Smallholding 282 29.4 14.2 8.1 7.1 Mid-holding 258 47.3 30.2 5.5 10.1 Males Landholding Category n Sons Servants Relatives Inmates Landless/Micro 366 23.5 1.1 3.8 14.2 Smallholding 282 36.5 4.3 5.4 16.3 Mid-Holding 258 65.9 17.4 5.5 31.0 Source: Cavan Sample Table 4. Spinning in each household by landholding category Mean per household Landholding category N/n Households with one or more spinners (%) Adult females Spinners Percent adult females spinners Non-artisan households (N) Landless/micro 514 53.9 1.43**** 0.76 55.6* Smallholding 562 53.7 1.65 0.83 49.4 Mid-holding 413 54.5 2.02**** 0.94 46.4 Weaving households (n) Landless/micro 366 53.8 1.43 0.76 55.7 Smallholding 282 52.5 1.77 0.78 44.8 Mid-holding 258 64.7 2.25 1.27 54.3 Source: Cavan Sample Difference from smallholders * p. < .10; **** p. < .001. Mann-Whitney test.